本文提出了一種創新的方法，利用人工智能和數據挖掘技術設計歷史模擬遊戲中的角色。我們專注於從影響力極大的歷史文本《三國演義》 (RTK) 中提取信息，以協助歷史模擬遊戲的開發。該研究利用自然語言處理 (NLP) 和命名實體識別 (NER) 從 RTK 文本中提取命名實體，形成知識圖譜數據集。機器學習基礎的節點重要性估計模型在這個數據集上進行訓練，以預測 RTK 中的角色重要性。我們將此模型的預測結果與參與者在調查中選擇的最愛角色進行對比，發現二者之間有很大的相似性。這也是一種新的框架，使歷史模擬遊戲的開發者可以基於中文文本預測角色的重要性。我們的結果突顯出這些方法在提高角色設計和敘事發展方面的潛力，從而提供了更沉浸式和吸引人的遊戲體驗。;This paper presents an innovative approach to character design in historical simulation games, leveraging artificial intelligence and data mining techniques. We focus on extracting information from an influential historical text, Romance of the Three Kingdoms(RTK) to assist in the development of historical simulation games. The study uses Natural Language Processing (NLP) and Named Entity Recognition (NER) to extract Named Entities from RTK text, forming a knowledge graph dataset. The machine learning-based node importance estimation model is trained on this dataset to predict character significance in RTK. We contrast the predictions of this model with the favorite characters selected by participants in a survey, discovering a strong similarity. This is also a new framework that enables developers of historical simulation games to predict the significance of characters based on Chinese text. Our results highlight the potential of these methodologies in enhancing character design and narrative development, thereby offering a more immersive and engaging gaming experience.
